Community Health Workers’ First-Session Alliance Strategies Drive Parent Engagement in Child Therapy

Behavioral parent training programs such as Parent–Child Interaction Therapy (PCIT) have decades of evidence behind them, yet more than half of eligible families never receive these services. Low-income families of color face the steepest barriers, from transportation and insurance gaps to stigma and a shortage of linguistically and culturally appropriate providers. A new study published in Frontiers in Child and Adolescent Psychiatry offers one of the first direct observational looks at how community health workers (CHWs) — trusted lay members of the communities they serve — keep these families engaged in treatment, and the findings point to a surprisingly simple lever: what happens in the very first session.

The research team, led by Hanan N. Salem of the University of California, Santa Barbara, together with collaborators at the University of Miami’s Mailman Center for Child Development and the community organization ConnectFamilias, examined a telehealth-delivered model known as PCIT + CHW. In this model, families receive standard PCIT from a therapist while also working with a CHW who helps them practice parenting skills at home, troubleshoot barriers to attendance and homework, and navigate practical challenges such as technology access, food and housing insecurity, and immigration-related needs. Previous trials of the model showed better clinical outcomes than standard PCIT among families recruited from underserved communities, but nobody had systematically measured what CHWs actually do in sessions to produce those gains.

To find out, the researchers turned to an existing observational instrument called the Therapist Alliance, Collaboration, and Empowerment Strategies (ACEs) Observational Coding System, a 12-item tool that rates how frequently and how thoroughly a provider deploys engagement strategies across four domains: alliance, collaboration, empowerment, and psychoeducation. Because the original tool was designed for therapists delivering structured evidence-based treatments, the team adapted it for the CHW context. They removed items that did not fit PCIT and added three strategies identified in prior qualitative work as distinctive to CHWs: leveraging shared characteristics, background, or lived experiences to build trust; meeting the caregiver where they are; and modeling advocacy when navigating systems such as schools or social services.

The modification also involved a technical change to the rating scale itself. The original version used a seven-point Likert scale in which adjacent values sometimes merged descriptors, for example with scores of 5 and 6 both indicating consistent and thorough use of a strategy. The researchers collapsed this into a four-point scale from 0 to 3, where 0 means the strategy was not used and 1 through 3 reflect increasing frequency, consistency, and thoroughness. Anchors were defined separately for each strategy; active listening, for instance, was operationalized as repeating or paraphrasing a parent’s perspective, summarizing multiple points the parent has made, or reflecting back the parent’s underlying feelings. The goal was to sharpen the distinctions between rating levels and thereby improve agreement between independent coders.

Three coders — two graduate students and one post-baccalaureate research assistant — completed an eight-week training involving 12 practice recordings and achieved at least 80 percent agreement with gold-standard codes before coding independently. The team then coded 52 recorded CHW sessions drawn from 26 families, capturing each family’s first CHW session and one randomly selected later session. The sample was overwhelmingly Latino: all 26 caregivers and 96 percent of the children were identified as Latino or Hispanic, half the families reported household incomes below the federal poverty line, and 88.5 percent of families received CHW sessions in Spanish. Children ranged from 2 to 8 years old, with an average age of about 4.8 years.

Not every adapted code survived scrutiny. Two strategies — leveraging shared experiences and modeling advocacy — appeared too rarely in the recorded sessions to analyze, and three items showed low interrater reliability as measured by intraclass correlation coefficients. Notably, two of those low-ICC items, offering suggestions about therapeutic tasks and providing psychoeducation about child services, actually showed high raw agreement between coders; the problem was a ceiling effect, with both strategies used so consistently that there was little variability to detect. The researchers retained them, yielding 11 final codes. Only the alliance composite and the total strategy score showed acceptable internal consistency, measured by McDonald’s omega, which the authors attribute to the relational nature of alliance strategies versus the more context-dependent character of collaboration, empowerment, and psychoeducation.

The descriptive results showed that CHWs used their full repertoire of strategies to a moderate extent on average, with alliance strategies at low-to-moderate levels. Two behaviors stood out as the most extensively delivered: providing psychoeducation about available child services and recognizing caregivers’ strengths and efforts to participate in therapy. The latter reflects PCIT’s skills-based structure, in which praise is itself a core technique, and CHWs are trained to model it by praising parents’ use of skills in session. Strategies such as addressing participation barriers, incorporating caregiver input, and psychoeducation about child behavior problems were used sparingly within the recorded sessions.

The regression analyses, though exploratory and uncorrected for multiple comparisons, produced a striking temporal pattern. Alliance strategies delivered during the first CHW session were significantly associated with caregivers’ perceived therapeutic alliance on the Working Alliance Inventory, their agreement that the tasks of therapy would help them achieve their goals, and their homework completion during the first phase of treatment. Individual first-session strategies also mattered: communicating positive regard and providing psychoeducation about child problems each predicted stronger alliance perceptions. By contrast, strategies measured at the second, randomly selected session showed no significant associations with relationship outcomes, suggesting that early alliance-building carries disproportionate weight — an extension to CHWs of a well-documented finding in therapist-client psychotherapy research.

Psychoeducation about child problems showed a more complicated profile. It predicted higher attendance rates both at the first session and when averaged across sessions, but its average use was negatively associated with caregivers’ acquisition of positive parenting skills. The authors offer a plausible mechanism: in PCIT, caregivers learn to attend to positive child behaviors, and even well-intentioned discussion of behavioral problems may unintentionally model attention toward those problems. They caution that the correlational design cannot establish direction — caregivers of children with more challenging behaviors may simply elicit more of these conversations — but the finding flags a training question about how CHWs frame discussions of child difficulties.

The study’s limitations are real: a small sample of 26 families and three CHWs, only two observed sessions per family, no correction for multiple comparisons, and a population limited to Latino caregivers all constrain generalizability. Much of the CHWs’ engagement work — advocacy, barrier-solving, bridging to services — likely happens outside formal sessions and went unobserved. Still, the authors argue this is the first study to use direct observational coding to evaluate engagement strategies in a CHW-delivered mental health model, and the implications for training are concrete: prepare CHWs to deploy alliance-based strategies, particularly active listening, partnership, and positive regard, from the very first contact, while augmenting the natural cultural and community knowledge that makes them effective in the first place.

Subject of Research: Community health worker engagement strategies in Parent–Child Interaction Therapy for low-income Latino families

Article Title: Modification of an observational tool to evaluate how community health workers engage low-income latino parents in parent–child interaction therapy

Article References: Salem, H. N., Rodriguez, F., Ison, J., Garcia, D., Jent, J. F., Davis, E. M., Alonso, B., Rothenberg, W. A., Espinosa, N., Perez, C., Rivera, Y. M., Orbano-Flores, L. M., Rincon, C., & Barnett, M. L. (2026). Modification of an observational tool to evaluate how community health workers engage low-income latino parents in parent–child interaction therapy. Frontiers in Child and Adolescent Psychiatry, 5, Article 1856265. https://doi.org/10.3389/frcha.2026.1856265

Image Credits: AI Generated

DOI: 10.3389/frcha.2026.1856265

Keywords: community health workers, Parent-Child Interaction Therapy, behavioral parent training, treatment engagement, Latino families, child mental health, observational coding, health disparities, Modification, observational, tool, evaluate